Unlike other HIV antivirals, such as protease inhibitors or non-nucleoside reverse transcriptase inhibitors, nucleoside reverse transcriptase inhibitors (NRTI) are pharmacologically inactive in their administered form and require phosphorylation by host cellular kinases to produce the active triphosphate metabolite. This triphosphate form resembles the naturally occurring deoxynucleotide triphosphate substrates of the viral reverse transcriptase and competes for HIV-1 RT binding and incorporation into viral DNA.
All NRTI s approved for the treatment of HIV, and the vast majority of all other NRTIs proposed in the patent or academic literature, lack a 3′-hydroxy function on the ribose moiety of the nucleoside. Examples include zidovudine (AZT), stavudine (d4T), lamivudine (3TC), zalcitabine (ddC), abacavir (ABC), didanosine (ddI) and tenofovir (TNF) (the latter being typically administered as the disoproxil fumarate prodrug). Upon phosphorylation, such a nucleoside or nucleotide analogue is covalently bonded by the reverse transcriptase enzyme to the nascent DNA strand, but the lack of a 3′-hydroxyl function in the nucleoside or nucleotide prevents further attachment of additional nucleotides. These NRTIs therefore terminate viral DNA strand prolongation, thereby leading to inhibition of HIV replication (Mitsuya et al 1990, Jacob Molina et al 1993, Reardon 1993).
The cornerstone of all current antiretroviral therapies (ART) is the use of NRTIs. NRTIs, however, are only able to retard HIV propagation in the blood stream and to date have been unable to eradicate HIV from patients. HIV operates by inserting its DNA into latent host cells involved in human immunologic memory. This mode of infection implies that patients are forced to take HIV antivirals lifelong in order to prevent the HIV titre from bouncing back after therapy has ended.
In practice, however, the effective administration period of a particular HIV drug for a given patient is dramatically limited by the emergence of “escape mutants.” An escape mutant is a virus that contains a discrete cluster of mutations that produces drug resistance and allows it to proliferate in the presence of the drug. Escape mutants arise in a patient due to the selective pressure of the particular antiviral(s) that the patient is taking. As a consequence, a drug's effective administration period is dependent on how quickly escape mutants arise and proliferate.
In countries consistently prescribing HIV antivirals it is becoming increasingly evident that the primary infection in new cases of HIV is often not with wild type HIV, but rather with a strain of HIV which is already partly or multiply resistant to the current antivirals. In other words, escape mutants which are generated in situ in infected patients can also be spread to naive patients by lateral or vertical transmission. This in turn means that even some patients who would otherwise be classified as treatment-naive are already infected with virus resistant to conventional first line therapies.
Multiple factors contribute to the selection of drug escape mutants including total HIV pool size, RT processivity and infidelity in viral genomic replication, viral fitness and multiple availabilities of target cells. By the late 1990s, evidence from long term use of combinations based on zidovudine (AZT) or stavudine (d4T) suggested that clusters of particular mutations in the RT were consistently generated. These mutation clusters are the prototype now known as Thymidine Analogue Mutations (TAMs). The presence of TAMs enhanced the likelihood of selecting further mutations and led to the development of more advanced NRTI resistance phenotypes that were not clearly within the family of thymidine analogues. Such phenotypes are now known as Nucleoside Analogue Mutation (NAM) and Multiple Drug Resistance (MDR) HIV.
Hypothesis for NRTI Resistance
AZT was the first antiretroviral to be widely used and not surprisingly was the first to generate escape mutants (Larder et al., 1989). However in view of the large number of mutations throughout the HIV genome in typical patient isolates it is not possible to produce the resistance phenotype in vitro using a recombinant RT enzyme bearing the particular TAM. As a consequence, the mechanisms through which TAMs confer resistance have not been straightforward to elucidate.
Various hypothetical models and theoretical predictions for the mechanism behind TAM resistance have been predicated on the involvement of nucleophilic attack by a pyrophosphate donor (Boyer et al, 2002 and Meyer et al, 2002). Presumably RT translocation theory is a key step in understanding the TAM associated resistance mechanism. This was, however, poorly understood until the end of 2002 because the RT pre- and post-translocation intermediates are transient and short-lived and not readily accessed experimentally.
The modern understanding of RT translocation theory holds that RT catalyzed DNA polymerization takes place in a detailed cascade fashion as illustrated in FIG. 3, which is adopted from Sarafianos et al (2003). These steps are                1) Binding of the DNA substrate by free enzyme E positions the 3′-primer end at the P-site (Primer site).        2) Binding of a dNTP close to the N-site (dNTP site) forms an “open” ternary complex.        3) A “closed” ternary complex is formed by enzyme conformational changes.        4) Phosphodiester bond formation between the 3′-OH primer terminus and the alpha phosphate of the dNTP is accompanied by release of pyrophosphate (PPi) to form the pre-translocated RT complex at the N-site.        5) Translocation of the primer terminus from the N-site to the P-site by forming a post-translocated complex which is a prerequisite for the next dNTP binding and continuation of DNA synthesis.        
If a DNA chain terminator nucleoside (NRTI) triphosphate (typically a nucleoside analogue which lacks a 3′-hydroxy function on the deoxyribose moiety) is used, it mimics its natural dNTP counterpart and binds to RT. After the analogous chemical processing, the incorporated NRTI forms a pre-translocation complex at the N-site of polymerization. This terminates further DNA synthesis due to the lack of a 3′-hydroxyl primer on the NRTI's deoxyribose moiety.
In contrast, TAM-related RT mutations employ a different nucleotide incorporation mechanism compared to wild type RT. Specifically, the new mechanism results in the release (excision) of the NRTI incorporated at the primer terminus, abrogating the chain terminating activity of the NRTI. This new mechanism is dependent on the interplay between the accumulation of complexes in pre-translocated states (at the N-site) and the availability of ATP or pyrophosphate donors, which are often abundant at the site of infection, i.e. normal lymphocytes.
ATP or pyrophosphate does not normally participate in viral DNA-polymerization reactions, but the structure of a RT expressing a TAM-related resistant phenotype facilitates their entry into a site adjacent to a newly incorporated NRTI. The equilibrium between pre- and post-translocational kinetic species provides a mechanism to ensure free access of the primer terminus to the N-site and also allows simultaneous binding of the pyrophosphate donor ATP at the P-site after the incorporation of the NRTI chain terminator and the release of pyrophosphate. When this occurs, ATP (or pyrophosphate) attacks the phosphodiester bond which links the incorporated NRTI at the end of the DNA, resulting in removal of the NRTI via pyrophosphorolysis. When the pyrophosphate donor is ATP, the NRTI is released as a dinucleoside tetraphosphate product. FIG. 4 illustrates this “primer rescue” in an AZT-terminated DNA (adopted from ClinicCareOptions™).
It is now believed that two distinctive mechanisms are involved in the phenotypic resistance to NRTI (Sluis-Cremer et al, 2000). The first, known as “primer rescue” activity, is described immediately above. Here, the chain-terminating nucleotide is removed from the 3′ end of the primer terminus through ATP-dependent or pyrophosphate-dependent pyrophosphorolysis. There is, however, another cluster of resistance phenotypes denoted as “discriminative mutants.” These mutants have an RT with enhanced ability to discriminate between NRTIs and native dNTPs. In this case, the mechanism leads to RT which is able to preferentially choose the right substrate (i.e. native dNTP), thereby avoiding chain termination by an NRTI and ensuring the propagation of the viral genome.
Generation of Mutations in HIV
Retroviruses such as HIV have the potential for rapid genetic diversification. While this is an energetically inefficient process, it offers clear adaptive advantages to the organism. The replication machinery used by HIV is particularly error prone, generates a large number of mutations and has the potential to lead to accumulation of mutations when the organism is under selective pressure.
Generally, the vast majority of mutations generated by viral replication result in less viable enzymes. Here, the accumulation of a second and especially a third mutation is less probable because the population pool for the less viable mutant, within which the second mutation must accumulate, will be diluted by the faster multiplying wild type organism.
Yet more viable viral mutants can arise and expand by two possible pathways. The first occurs when there is rapid outgrowth of a highly resistant variant that is already present in the overall viral population. Most frequently this is a single point mutation that confers phenotypic resistance to a selective pressure. In the context of drug escape mutations examples include K103 rapidly induced by the non-nucleoside reverse transcriptase inhibitor nevirapine.
The second pathway occurs when there is continued viral replication in the presence of selective pressure. This allows the progressive accumulation of mutations that can then be expanded. In this case, the probability of mutation accumulation is related to the amount of virus replication that is occurring. That is, at higher viral loads (e.g. >200,000 copies/ml), accumulations of double mutations can occur. Accumulation of triple mutations, however, are rare and can only result as a consequence of a complex therapeutic regimen, typically involving several different drugs, that is challenging for the patient to adhere to. It is therefore extremely difficult for even a diligent patient to ensure that all active ingredients are present in the blood at levels above the necessary inhibitory concentrations over the full 24 hour period of each day “24 hour trough level”. Here, temporary removal of any one of the selective pressures of drug treatment due to lapses in the administration/24 hour trough level of one or more drugs allows unbridled viral replication, thereby permitting the generation and establishment of many new mutants. When the selective pressure is once again applied (i.e. resumption of complex drug therapy), the few new mutants that have accumulated another point mutation which confers better drug resistance can expand in a manner similar to that seen for the first pathway (see above).
The discussion above focuses on accumulation of point mutations as opposed to, for example, deletion or addition mutations. Here, however, a scenario similar to that described for a triple mutation is applicable. That is, most deletion/addition mutations initially involve a single nucleotide. This has the effect of completely altering the downstream amino acid sequence of the encoded protein if the change occurs within the coding region and leads to a truncated and/or inactive protein. In order to preserve the reading frame and to alter the final protein by the deletion or addition of one single amino acid, three nucleotides must be deleted/added. Since inactive enzymes reduce the viability of an HIV organism, particularly if the enzyme affected is RT, the deletion/additions will not accumulate per se, but must occur simultaneously. In other words the equivalent of a triple mutation must occur in a single event, which is highly uncommon (see Boyer et al (2004) J Virol 78(18):9987-9997, which is hereby incorporated by reference in its entirety).
As a consequence of this process for triple mutant accumulation/introduction, it was not until relatively recently that HIV virus exhibiting at least three mutations in RT that creates particularly potent resistance to multiple drugs became established. For example, in the United States it was 1992 when the FDA approved the use of combination drug therapy (ddC and AZT). Yet it was not until September of 1995 that clinical trials showed that the combination of AZT with ddC or ddI was more effective than AZT alone. It has only been as a result of the use of combination therapies, where multiple drugs are employed, but in dosage regimes effectively unable to guarantee an adequate 24 hour trough level of the respective drugs, that the particularly problematic strains of multiresistant HIV virus known in the Western world today have been generated.
Primer Rescue Mutations
The TAM primer rescue mutant originally described comprised various permutations within a group of six drug resistant phenotypes at amino acid positions M41L, D67N, K70R, L210W, T215Y/F and K219Q/E on RT (Larder and Kemp, 1989, Schinazi et al, 2000). Early data pointed to two distinctive mutational pathways for the development of multiple TAM primer rescue mutants, both occurring by unknown factors. The first pathway resulted in an amino acid substitution at codon 210 (210W) and was preferentially associated with mutations at codons 41 (41 L; greater than 98%) and 215 (215Y; greater than 94%) as well as a substitution at codon 67 (67N). The second pathway generated a mutation at codon 219 (219K/E), which was preferentially associated with mutations at codons 67 (67N) and 70 (70R)(Yahi et al, 1999). There were therefore two phenotypic patterns: (1) L210W, M41L, T215Y/F, ±D67N, which conferred high levels of viral resistance to AZT and d4T and (2) K219K/E, D67N, K70R, which conferred moderate levels of viral resistance to AZT and d4T.
Marcelin et al (2004) summarized the prevalence of TAM primer rescue-related mutations in virologic failure pateints. Here, 1098 RT sequences were investigated and gave two genotypic patterns as indicated in FIG. 1 and FIG. 2. While different genetic backgrounds may have been present prior to therapy, the sequence and composition of the antiretroviral therapy undertaken when combined with individual differences in pharmacology resulted in viral resistance not only to AZT and d4T but also to other NRTIs. Depending on the mutational pattern present, drug resistance included abacavir (ABC), didanosine (ddI), tenofovir (TNF), lamivudine (3TC), emtricitabine (FTC) and zalcitabine (ddC). Hence, the emergence of primer rescue-related TAMs often plays an important role in the further development of more pronouncedly resistant HIV genotypic patterns. Therefore, one step in preventing multiple nucleoside resistance is to develop a new NRTI with the goal of avoiding the accumulation of primer rescue related TAMs.
Primer rescue-related TAM mutations can evolve concomitantly with other families of escape mutants that typically emerge from combination antiretroviral therapy (otherwise known as cocktail therapy). Today, the cocktail “combivir” (AZT+3TC) is the most frequently used and recommended first line therapy regimen for treatment of naïve HIV patients. It leads, however, to escape mutants which are resistant to both drugs. For example, Miller et al (1998) reported that 3TC-resistant virus with an M184V mutation was selected just 4-12 weeks after initiation of AZT+3TC combination therapy. In time, additional AZT-associated mutations gradually emerged, giving a characteristic genotypic pattern of M184V, M41L, D67N, K70R, L210W, T215Y/F and K219Q/E which is commonly found in treatment experienced patients today. Additional mutations in RT at positions H208, R211, and L214 (Sturmer et al, 2003) and at position G333 (Kemp et al 1998) are reported to be involved in AZT-3TC double resistance and, in particular, to confer an increase in the ability to resist AZT. Therefore, the genotypic context of primer rescue related TAMs has been expanded to include permutations within M184V, M41L, D67N, K70R, H208Y, L210W, R211K, L214F, T215Y/F, K219Q/E and G333E.
Other types of mutations generally seen in treatment experienced patients are V118I and E44D/A. These mutations are strongly correlated to prior exposure to ddI and d4T. In addition, they are often associated with the presence of specific TAM clusters including M41L plus T215Y/F or D67N plus L210W. The result is increased primer rescue-related TAM resistance to the family of thymidine analogues as well as a distinctive role in the dual resistant to AZT+3TC (Montes et al, 2002, Girouard et al, 2003).
The prevalence of drug escape mutants increases as a function of the number of NRTIs used during the course of therapy and forms a pattern of expanded TAMs or NAMs comprising various permutations within M41L, E44D/A, D67N, K70R, V118 μM184V, H208Y, L210W, R211K, L214F, T215Y/F, K219Q/E and G333E. This cluster is also commonly refractory to AZT- and d4T-containing combination therapies and cross-resistant to the entire class of NRTIs.
Significant resistance to thymidine analogues, notably AZT, d4T and TNF, is also found in escape mutants having an amino acid deletion at position 67(▴67) in the finger region of RT often in association with an amino acid substitution at T69G concomitant with TAM (see Imamichi et al 2000 and 2001). An enhanced RT polymerization activity, which is associated with this particular genotype, is proposed to result in more efficient pyrophosphorolysis-dependent primer excision (described above), leading to the increased resistance Boyer et al, (2004) have also observed that ▴67 concomitant with TAM conferred an increased ability to facilitate primer rescue (excision) viral resistance to AZT and to TNF as compared to TAM alone.
HIV is co-evolving as antiretroviral therapy develops. New mutation phenotypes emerged when double- and triple-nucleoside analogue cocktails were employed in the clinical management of HIV, especially in treatment-naive patients. Complex therapeutic regimens, requiring multiple drugs taken at various times during the day, some with and some without food, are challenging for patients. Failure to comply exactly with these dosing regimes leading to 24 hour trough failures have facilitated the emergence of multiple NRTI resistant HIV viruses, predominantly as a result of virus acquired NAMs or MDRs. For example, a number of groups (e.g. Mas et al, 2000) have observed the emergence of the mutant T69S-XX virus associated with AZT use. This mutant, has a 6-bp insertion in the coding region of its RT between the nucleic acids specifying amino acids 69 and 70. The resulting double amino acid insertion complexes (typically SS, SG or AG insertions) not only led to viral resistance to AZT but also to nearly the entire collection of NRTIs including d4T, 3TC, ddI, ddC and ABC, and TNF. An enhanced pyrophosphorolysis-dependent primer rescue is seen with the T69S+double amino acid insertion, particularly in the presence of TAMs. This phenomenon is typically associated with the “M41L/T215Y” or “M41L/L210W/R211K/L214F/T215Y” resistant phenotypes and plays an important phenotypic role in multiple nucleoside resistance (Meyer et al, 2003).
Another class of MDR has an amino acid substitution at codon Q151M. This mutation is observed at a relatively low frequency in the clinic and often presents together with secondary mutations of A62V, V75I, F77L and F116Y. It confers, however, a significant resistance to nearly the entire class of NRTIs. In addition, it has been observed associated with TAMs, typically the “M41L, L210W and T215Y/F” or “D67N, K70R and K219K/E” genotypes. It emerges in patients that have experienced heavy treatment with AZT/ddI and AZT/ddC combination regimens.
L74V is most frequently selected by ddI monotherapy (Martin et al, 1993) and displays cross-resistance to ABC and 3TC. Its effect on producing viral escapes is dependent upon the presence of other mutations. Resistance surveys suggest that the frequency of L74V is linked significantly with TAM, typically in an M41L, L210W and T215Y/F background (Marcelin et al, 2004) even though the L74V mutation was thought to cause a diminution effect in viral replication and to resensitize AZT-resistant viruses that contain a number of TAMs (St. Clair et al, 1991). A combination of the L74V and M184V mutations in HIV-1 RT is the most frequent pattern associated with resistance to both ABC and ddI (Harrigan et al, 2000 and Miller et al, 2000).
Although high-level resistance to ABC typically requires multiple mutations comprising K65R, L74V, Y115F and M184V, a single mutation, M184V, often emerges first. This mutation, now recognized as a key mutation in the discriminant mechanism of drug escape resistance, confers a moderate decrease in ABC susceptibility (Tisdale et al, 1997). A CNA3005 study in which a total of 562 patients randomly received AZT and 3TC with either ABC or ddI, showed a slow but steady increase in the proportion of patients carrying a TAM in the AZT and 3TC plus ABC arm. By week 48, up to 56% of the patients had at least one primer rescue-related TAM (1xTAM) over and above the rapidly induced M184V mutation (Melby et al, 2001), illustrating the importance of preventing the emergence of primer rescue-related resistance. Similarly, in vitro passage of AZT-resistant virus bearing the genotypic pattern of 67, 70, 215 and 219 under 3TC selective pressure resulted in the selection of the M184V mutation and conferred cross-resistance to ABC (Tisdale et al, 1997). This again highlights the concept that treating the pre-existing of primer rescue-related TAM and preventing the accumulation of primer rescue-related mutants is a pivotal step in avoiding development of multiple nucleoside resistance.
It has become increasingly clear that the K65R mutation quickly appears in a very high proportion of patients who are receiving TNF or ABC. Valer et al (2004) reported that K65R increased in prevalence in their Madrid hospital from <1% between 1997-2000 to 7% in 2003 and 12% in the first 4 months of 2004. The effect of the K65R mutant is exacerbated in the presence of other mutations associated with decreased susceptibility to ABC, 3TC, ddI and ddC (Parikh et al, 2003). Yet the simultaneous appearance of K65R of primer rescue-related TAM genotypes, although rarely occurring, leads to a more profound effect on the primer rescue (excision) of TNF than of AZT (Naeger et al, 2001). TNF was reported to be active against HIV-1 with up to 3xTAMs unless the TAM cluster included an M41L or L210W mutation. Currently it is unclear why TAMs could reverse some of the effects of K65R, which is otherwise thought to impede primer excision mutants with respect to susceptibility to TNF and ABC.
Finally, the T69D mutation was initially identified for its role in causing ddC resistance. It has also been reported to be associated with a decreased response to ddI when it occurs in combination with the T215Y mutation and other of primer rescue-related TAM genotypes.
For many years the WHO and DHHS (US Department of Health and Human Health Service) have recommended first-line antiretroviral therapy on treatment naïve patients consisting of administering d4T or AZT in combination with 3TC plus nevirapine or efavirenz (Guidelines for the Use of Antiviral Retroviral Agents in HIV-1-Infected Adults and Adolescents, Jul. 14, 2003 and Mar. 23, 2004). A substantial number of HIV-infected patients have, however, experienced treatment failure while on their initial highly active antiretroviral therapy (HAART) regimens, suggesting that these patients are already infected with drug escape viruses. Primer rescue-related TAM resistance mutants continue to play a pivotal role in the development of drug resistance. Thus the development of drugs or therapeutic methods that counteract the effect of primer rescue-related TAM resistance mutants could potentiate or prolong the use of existing NRTIs for treating treatment-naïve patients and could also be used to treat the primer rescue-related resistance mutant-carrying HIV infected population in a salvage therapy.
Drug Strategies for Preventing/Inhibiting Primer-Rescue Mutants
Primer rescue and discriminative mutations often appear together in the same mutant genotype, largely due to current therapeutic strategy. A M184V mutation is representative of the family of discriminative mutants. If, however, it occurs in conjunction with primer rescue-related mutants such as M41L, D67N, K70R, L210W, T215Y/F, and K219Q/E, it plays a role in the dual resistance to AZT and 3TC (Miller et al., 1998).
These primer rescue and discriminative resistance phenotypes seem to correlate with different clusters of mutations in RT. For example, AZT-associated mutations comprising various permutations within M41L, E44D/A, D67N, K70R, V118I, M184V, H208Y, L210W, R211K, L214F, T215Y/F, K219Q/E and G333E, an MDR T69S mutation with 6-bp insertions and a ▴67 typically exhibit primer rescue mutant activities. On the other hand, mutations at positions 65, 74, 89, 151, and 184 lead to the ability to discriminate between NRTIs and the respective dNTP counterparts or they may be involved in the repositioning of the primer-template complex.
In the recent article “Designing anti-AIDS drugs targeting the major mechanism of HIV-1 RT resistance to nucleoside analog drugs” (IJBCB 36 (2004) 1706-1715, which is hereby incorporated by reference in its entirety), Sarafianos et al conclude that the primer rescue (excision) mechanism could only occur before RT translocation at the N-site and further conclude that it has become the dominant mechanism of NRTI resistance. In the chapter entitled “Strategies for Inhibition of the Excision Reaction” (see page 1711), they propose three approaches to defeat such a resistance mechanism:                1. use of new antivirals that interfere with the productive binding of ATP (at the P site), presumably by binding at or near the ATP-binding site, thereby blocking the excision reaction without affecting the forward reaction of DNA synthesis.        2. use of compounds that can block DNA synthesis but are somehow resistant to excision, such as borano- or thio-substituted alpha phosphate variants of the current NRTIs. Similarly, variants of the current NRTIs can be engineered to reposition the extended/terminated template/primer in a non-excisable mode, as suggested by the poor excision capacity of the M184I/V mutants induced by 3TC.        3. use of dinucleotide tetraphosphate based inhibitors to provide bi-dentate binding at both N- and P-sites.        
Each of these three proposed approaches to preventing primer rescue mechanisms of NRTI resistance is open to criticism for various theoretical shortcomings. For example, in the first approach ATP binding is not required for normal RT functions. Thus, countermeasures based on inhibiting ATP or pyrophosphate binding by competition or blockage will not prevent resistance development because the fitness of the underlying virus will not be compromised by such agents. In other words, resistance mutations will arise at no evolutionary cost. The abundant amount of ATP present in normal lymphocytes also challenges the rationale behind this approach.
In the second proposed approach, it seems likely that borano- or thio-substituted alpha phosphate analogues would select for the discriminative resistant mutants, as has been seen with 3TC and FTC, and produce HIV resistance mutants.
The third proposed approach is limited by the need for pharmacokinetic uptake into the target cell of the large and highly charged tetraphosphate dinucleotide species. This will be a severe pharmaceutical and drug delivery challenge.
It is noteworthy that each of Serafaniano's approaches, including approach 1 which is not antiviral in itself, but presupposes co-administration of a conventional NRTI, is based on variants of the current generation of NRTIs. That is, compounds that lack a 3-hydroxyl function and therefore act as obligate chain terminators.
In contrast to the “classic” NRTIs discussed above (i.e. those lacking a 3′-hydroxy function), Ohrui et al (J Med Chem (2000) 43, 4516-4525, which is hereby incorporated by reference in its entirety) describe 4′-C-ethynyl HIV inhibitors:

These compounds retain the 3′-hydroxy function but nevertheless exhibit activity against HIV-1, including a typical discriminative MDR strain bearing the A62V, V75L, F77L, F116Y and Q151 M mutations. The mechanism of action was postulated to be through affinity to the nucleoside phosphorylating kinase. It was, however, also observed that these compounds may be functioning as DNA chain terminators due to their neopentyl alcohol character and the severe steric hindrance of the vicinal cis 4′ substituent, which resulted in a sharply diminished reactivity of the 3′-hydroxy.
Kodama et al (Antimicrob Agents Chemother (2001) 1539-1546, which is hereby incorporated by reference in its entirety) describe a very similar set of compounds bearing a 4′-C-ethynyl group adjacent to the retained 3′-hydroxy function that were assayed in cell culture with additional HIV resistant strains. Since Kodama et al did not prepare the triphosphates of their compounds, they were unable to elucidate the mechanism of action but infer from various circumstantial observations that the compounds are indeed acting as NRTIs. Kodama et al later reported (abstract 388-T, 2003 9th Conference on Retroviruses and Opportunistic Infections, which is hereby incorporated by reference in its entirety) that under the selective pressure of their 4-C-ethynyl nucleoside in vitro, breakthrough resistant HIV bearing T1651 and M184V mutations located in the RT catalytic site were found. This mutant phenotype is manifestly a discriminative type of mutation and is heavily cross resistant to 3TC. Steric conflict blocking 4-C-ethynyl nucleoside incorporation was thus implicated. This has been established with the 3TC inhibitory mechanism and therefore almost certainly represents the discriminative resistant mechanism. It therefore seems unlikely that the Kodama compounds will provide guidance in addressing the mutants facilitating primer rescue (ATP or pyrophosphate mediated excision).
Chen et al (Biochemistry (1993) 32:6000-6002, which is hereby incorporated by reference in its entirety) conducted extensive mechanistic investigations on a structurally related series of compounds bearing an azido group at 4′:

Chen demonstrated that RT efficiently incorporates two consecutive 4′-azidothymidine monophosphate nucleotides, which terminates chain elongation. In addition, RT was also able to incorporate a first 4′-azidothymidine monophosphate, followed by a native dNTP and a then a second 4′-azidothymidine nucleotide, which also led to chain termination. Note that both of these mechanisms resulted in a 4′-azidothymidine monophosphate residing at the terminated DNA primer terminus, which is an inhibitory mechanism very reminiscent of the current NRTIs. It was also apparent that the cellular (ie non-viral) polymerases a and B were each able to incorporate a single 4′-azido nucleotide, but not a second, into the nascent chain of the host DNA. These cellular polymerases then allowed the host DNA chain to elongate with further native dNTPs and so permanently incorporated the NRTI nucleotide into host DNA genes. These compounds have not been pursued in humans because misincorporation of non-native nucleotides by cellular enzymes has clear implications in carcinogenesis. Similarly, the pharmaceutical development of the Kodama corresponding 4′-C-ethynyl compounds was stopped, allegedly due to severe toxicity in higher organisms.
EP 341 911 describes an extensive family of 3′-C-hydroxymethyl nucleosides of the formula

and proposes their use predominantly against herpesviruses such as CMV, but also against retroviruses. WO92/06201 also discloses a similar set of compounds and indications.
U.S. Pat. No. 5,612,319 (which is hereby incorporated by reference in its entirety) discloses the retroviral activity of 2′-3′ dideoxy-3′-C-hydroxymethylcytidine against wild type HIV-1IIIB and the simian equivalent, SIV-1, in an acute cynomolgus monkey model of HIV infection. This publication proposes the use of the compound as a post-exposure prophylaxis agent, especially against needle-stick injuries. Post exposure prophylaxis implies that the active ingredient is immediately administered to people such as medical personnel, who have unwittingly jabbed themselves with a potentially HIV-infected syringe. In order to ensure rapid treatment of an understandably shocked health care professional, a self administered spring-loaded syringe, such as are used for antidotes to chemical and biological warfare, is a preferred administration route.
The intention of post-exposure prophylaxis is to prevent the infection from establishing itself rather than treating an on-going infection. As such, it was intended that treatment was to be carried out for a short time period such as 24-48 hours, using extremely high doses of the compound. This publication states that because of the discrete time period of administration, transient toxicity is acceptable because one is trying to prevent an incurable disease. The post-exposure prophylactic method described in U.S. Pat. No. 5,612,319 has never been tried in humans—indeed to our knowledge 2′-3′ dideoxy-3′-C-hydroxymethylcytidine has not been administered to humans at all.
In 1994 when the application granting as U.S. Pat. No. 5,612,319 was filed, multi-resistant HIV as it is known today had not arisen in any cogent form. Today's multi-resistant HIV has primer rescue mutations induced by and accumulated from many years of selective pressure from NRTI therapy. In other words, the HIV and especially the RT existent at the time these patents were granted was structurally and mechanistically very different from today's viruses.